Human-robot collaborative transport personalization via Dynamic Movement Primitives and velocity scaling

TL;DR

The paper addresses safe and personalized human-robot collaboration during cooperative transport tasks by combining offline DMP-based trajectory personalization with online velocity scaling driven by interaction forces. It models trajectories with $\ddot{y}(t)=\alpha_y(\beta_y(g-y(t))-\dot{y}(t))+f_{DMP}$ and uses a forcing term $f_{DMP}=\frac{\sum_{i=1}^{N}\Psi_i(x)\omega_i}{\sum_{i=1}^{N}\Psi_i(x)} x (g-y_0)$ to reproduce demonstrations, enabling generalization to new start/goal states. The method decouples planning from execution, allowing collision checks offline while enabling force-driven speed adjustments during execution. Experiments with 12 participants comparing DMP-based methods to a BiTRRT planner show higher subjective fluency and trust for DMPs, with physiological data indicating lower cognitive load; velocity scaling further enhances perceived cooperation, particularly when the user can modulate speed. The results suggest that personalization via DMPs and force-based velocity scaling can substantially improve HRI in industrial co-manipulation tasks.

Abstract

Nowadays, industries are showing a growing interest in human-robot collaboration, particularly for shared tasks. This requires intelligent strategies to plan a robot's motions, considering both task constraints and human-specific factors such as height and movement preferences. This work introduces a novel approach to generate personalized trajectories using Dynamic Movement Primitives (DMPs), enhanced with real-time velocity scaling based on human feedback. The method was rigorously tested in industrial-grade experiments, focusing on the collaborative transport of an engine cowl lip section. Comparative analysis between DMP-generated trajectories and a state-of-the-art motion planner (BiTRRT) highlights their adaptability combined with velocity scaling. Subjective user feedback further demonstrates a clear preference for DMP- based interactions. Objective evaluations, including physiological measurements from brain and skin activity, reinforce these findings, showcasing the advantages of DMPs in enhancing human-robot interaction and improving user experience.

Figures (13)

• Figure 1: Participant performing the collaborative transport of a section of the engine cowl lip. The subject is wearing physiological sensors used to evaluate the stress level during transport.
• Figure 2: The velocity $v$ and force $f$ vectors, the versor $\hat{t}$, and the vector of force components along the velocity direction.
• Figure 3: Comparison between aligned wrench and speed override.
• Figure 4: Ratio between average trajectory height and users' heights, for each user in the study.
• Figure 6: Normalized scale scores for different trajectory generation methods across various human-robot interaction metrics.